Observations of the large earth bumblebee, Bombus terrestris (L.), in native vegetation were collated to determine the extent to which this exotic species has invaded Tasmanian native vegetation during the first 9 years after its introduction. The range of B. terrestris now encompasses all of Tasmania's major vegetation types, altitudes from sea level to 1260m a.s.L, and the entire breadth of annual precipitation in the state from more than 3200 mm to less than 600 mm. Observations of workers carrying pollen, together with the presence of large numbers of bumblebees at many localities across this range indicate that colonies are frequently established in native vegetation. Evidence that colonies are often successful was obtained from repeated observations of the species during more than 1 year at particular sites. Unequivocal evidence of colonies was obtained from six National Parks, including four of the five in the Tasmanian Wilderness World Heritage Area (WHA). Indeed, the species has been present in the WHA for at least as long as it has in the city of Hobart, where it was first recorded. In southwestern Tasmania, evidence of colonies was obtained up to 40km from gardens, 61 km from small towns and 93 km from large towns. Hence, contrary to previous suggestions, the species is established in the most remote parts of Tasmania and is not dependent on introduced garden plants. Given their strong record of invasion, it is likely that B. terrestris will form feral populations on the mainland of Australia and in many other parts of the world if introduced. Because of their likely negative impacts on native animals and plants, and potential to enhance seed production in weeds, the spread of bumblebees should be avoided.

Summary1. Ecosystem change is predicted to become more prevalent with climate change. Widespread dieback of cushion plants and bryophytes in alpine fellfield on Macquarie Island may represent such change. Loss of the keystone endemic cushion plant, Azorella macquariensis, was so severe that it has been declared critically endangered. 2. We document the dieback and its extent. Due to the rapidity of the event, we sought to infer causes by testing two mechanistic hypotheses: (i) that extensive dieback was due to a pathogen and (ii) that dieback was a consequence of a change in climate that induced stress in several susceptible species. We searched for pathogens using both conventional and next-generation sequencing techniques. We examined plant functional morphology in conjunction with a long-term climate record of plant-relevant climate parameters to determine whether environmental conditions had become inimical for A. macquariensis. 3. Dieback was found across the entire range of A. macquariensis. A survey found 88% of 115 stratified/ random sites contained affected cushions and 84% contained dead bryophytes. Within-site dieback increased over time. 4. No conclusive evidence that A. macquariensis deaths were caused by a definitive diseasecausing pathogen emerged. However, the presence of bacterial, fungal and oomycete taxa, some potentially pathogenic, suggested that stressed cushions could become susceptible to infection. 5. The primary cause of collapse is suspected failure of A. macquariensis and other fellfield species to withstand recent decadal changes in summer water availability. Increased wind speed, sunshine hours and evapotranspiration resulted in accumulated deficits of plant available water spanning 17 years (1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008). High vulnerability to interrupted water supply was consistent with functional morphology of A. macquariensis, and climate change has altered the species' environment from wet and misty to one subject to periods of drying. Journal of Applied Ecology 2015Ecology , 52, 774-783 doi: 10.1111Ecology /1365Ecology -2664 were complex and multiple stressors appeared to be impacting cumulatively may be relevant to other locations.

Summary Australia's sub‐Antarctic Macquarie Island is presently undergoing one of the most ambitious vertebrate pest eradication programmes ever undertaken. The anticipated success of this programme will release the island's tundra‐like vegetation from well over a century of grazing and disturbance from House Mouse (Mus musculus), Ship Rat (Rattus rattus) and most significantly European Rabbit (Oryctolagus cuniculus). This study describes results from 30 years of vegetation quadrat monitoring (prior to the most recent and comprehensive pest eradication programme) when lower level pest animal control programmes were underway. Plant species were assigned to one of five distinct functional plant groups: Indigenous short‐lived perennials, Introduced short‐lived perennials, Indigenous perennials rarely grazed by rabbits, Indigenous perennials occasionally grazed by rabbits and Indigenous long‐lived perennials heavily grazed by rabbits, with one species, Agrostis magellanica, analysed as a sixth monospecific group. Results from monthly rabbit counts were used to compare changes in abundance of these six groups under different rabbit populations. It was found that there were three distinct phases of rabbit activity during the study period, indicated by (i) an initial very high count year in 1980–1981, followed by (ii) 20 years of low counts ending in 2001–2002 after which (iii) counts rose to medium/high until the commencement of the eradication programme in 2010–2011. Vegetation composition and progression were distinct for these three rabbit count phases. The first four of the plant functional groups decreased under lower count periods and increased in cover under higher rabbit count periods. Agrostis magellanica appears to respond primarily to interspecies competition and is disadvantaged under extended periods of low rabbit numbers. Indigenous long‐lived perennials heavily grazed by rabbits, which includes the large tussocks and megaherbs, is inversely related to rabbit numbers. During the study period, there has also been an overall decline in plant species richness with average species count per quadrat falling by between 0.6 and 2.7 taxa. This study attempts to address the observed vegetation change from this long‐term monitoring, to discuss other potential contributing factors and to use the results to predict likely future vegetation changes after eradication of vertebrate pests.

Abstract.A decline was observed in the subantarctic Macquarie Island endemic cushion, Azorella macquariensis, during the summer of 2008-2009, resulting in the listing of the species as critically endangered in 2010. Photographs of A. macquariensis in the period 2009-2013 were used to (1) identify types of damage, (2) determine the likely causes of three distinct types of damage, (3) establish whether dieback was spreading from affected to unaffected sites and (4) find out whether dieback was associated with the expansion of Agrostis magellanica. Grey damage occurred on the most windexposed parts of cushions and on the most wind-exposed sites. Speck damage occurred in the opposite situations and was consistent in its location, attributes and timing with rabbit grazing. Yellow dieback was sporadic in its occurrence. Its symptoms were consistent with those of a pathogen. Yellow damage expanded between spring 2009 and autumn 2010, with neither grey nor speck damage increasing. Yellow damage was associated with a marked decline in live cushion cover in plots between 2010 and 2013. The cushion was not eliminated from any plots, despite increased cover of A. magellanica in plots with dead cushions. Only one site not affected by yellow damage in 2010 had become affected by 2013. Given these results, and given that yellow damage has been observed in the past, 2008-2010 may have been an infrequent extreme outbreak of a pathogen and/or a response of a pathogen to ongoing climatic change.

ContextFire management advice for fauna conservation in protected areas must often be based on expert opinion and extrapolation from very few scientific studies. More monitoring and research are needed to better inform land managers tasked with both managing for biodiversity and managing the threat of bushfires. AimsTo document changes in the activity of native and introduced mammal species in response to planned burns in a small, isolated woodland reserve in Tasmania. MethodsOver a 10-year period, mammal activity was monitored before and after two separate moderate-intensity planned burns in 20-ha management blocks and in unburnt blocks of similar size by using live-trapping and camera-trapping. Vegetation density was monitored concurrently, and we also searched burnt areas for animals killed by fire. Key resultsThe activity of most mammal populations was largely unaffected by the two planned burns. However, during one of the burns, over 20 rufous-bellied pademelons were directly killed as a result of a need to conduct a backburn. The population recovered after 3 years. The activity of red-necked wallabies, common brushtail possums and short-beaked echidnas generally increased across the whole study area during the 10-year monitoring period. Limited evidence suggests that eastern barred bandicoot and European rabbit activity increased after fire. No swamp rat activity was recorded in burnt areas following the planned burns. Unexpectedly they did not recolonise burnt areas and also ceased to be active in control areas for the last 3 years of the study; we hypothesise that this may be due to the increased dryness and thinning of vegetation. ConclusionsWe found that most of the mammal populations within this small, isolated reserve were resilient to the planned burning program, with no or limited short-term effect for all but one species. The absence of swamp rats from burnt or unburnt areas for the last 3 years of our study suggests that factors other than fire are also affecting this species. ImplicationsPlanned burning is an important tool for biodiversity conservation, but its use needs to be underpinned by empirical data because mammal fire responses are likely to be site-, time- and context-specific.